--- /home/cpdev/src/classpath/gnu/xml/transform/StreamSerializer.java 2005-07-10 05:32:35.000000000 +0000 +++ gnu/xml/transform/StreamSerializer.java 2005-06-30 05:34:21.000000000 +0000 @@ -41,14 +41,9 @@ import java.io.IOException; import java.io.OutputStream; import java.io.UnsupportedEncodingException; -import java.nio.ByteBuffer; -import java.nio.CharBuffer; -import java.nio.charset.Charset; -import java.nio.charset.CharsetEncoder; import java.util.Collection; import java.util.Collections; import java.util.HashMap; -import java.util.HashSet; import java.util.Map; import javax.xml.XMLConstants; import org.w3c.dom.Attr; @@ -73,85 +68,8 @@ static final int KET = 0x3e; // > static final int EQ = 0x3d; // = - /** - * HTML 4.01 boolean attributes - */ - static final Map HTML_BOOLEAN_ATTRIBUTES = new HashMap(); - static - { - HashSet set; - - set = new HashSet(); - set.add("nohref"); - HTML_BOOLEAN_ATTRIBUTES.put("area", set); - - set = new HashSet(); - set.add("ismap"); - HTML_BOOLEAN_ATTRIBUTES.put("img", set); - - set = new HashSet(); - set.add("declare"); - HTML_BOOLEAN_ATTRIBUTES.put("object", set); - - set = new HashSet(); - set.add("noshade"); - HTML_BOOLEAN_ATTRIBUTES.put("hr", set); - - set = new HashSet(); - set.add("compact"); - HTML_BOOLEAN_ATTRIBUTES.put("dl", set); - HTML_BOOLEAN_ATTRIBUTES.put("ol", set); - HTML_BOOLEAN_ATTRIBUTES.put("ul", set); - HTML_BOOLEAN_ATTRIBUTES.put("dir", set); - HTML_BOOLEAN_ATTRIBUTES.put("menu", set); - - set = new HashSet(); - set.add("checked"); - set.add("disabled"); - set.add("readonly"); - set.add("ismap"); - HTML_BOOLEAN_ATTRIBUTES.put("input", set); - - set = new HashSet(); - set.add("multiple"); - set.add("disabled"); - HTML_BOOLEAN_ATTRIBUTES.put("select", set); - - set = new HashSet(); - set.add("disabled"); - HTML_BOOLEAN_ATTRIBUTES.put("optgroup", set); - - set = new HashSet(); - set.add("selected"); - set.add("disabled"); - HTML_BOOLEAN_ATTRIBUTES.put("option", set); - - set = new HashSet(); - set.add("disabled"); - set.add("readonly"); - HTML_BOOLEAN_ATTRIBUTES.put("textarea", set); - - set = new HashSet(); - set.add("disabled"); - HTML_BOOLEAN_ATTRIBUTES.put("button", set); - - set = new HashSet(); - set.add("nowrap"); - HTML_BOOLEAN_ATTRIBUTES.put("th", set); - HTML_BOOLEAN_ATTRIBUTES.put("td", set); - - set = new HashSet(); - set.add("noresize"); - HTML_BOOLEAN_ATTRIBUTES.put("frame", set); - - set = new HashSet(); - set.add("defer"); - HTML_BOOLEAN_ATTRIBUTES.put("script", set); - } - - protected final String encoding; - final Charset charset; - final CharsetEncoder encoder; + protected String encoding; + boolean compatibilityMode; final int mode; final Map namespaces; protected String eol; @@ -178,8 +96,16 @@ encoding = "UTF-8"; } this.encoding = encoding.intern(); - charset = Charset.forName(this.encoding); - encoder = charset.newEncoder(); + compatibilityMode = true; + if (encoding.length() > 3) + { + String p = encoding.substring(0, 3); + if (p.equalsIgnoreCase("UTF") || + p.equalsIgnoreCase("UCS")) + { + compatibilityMode = false; + } + } this.eol = (eol != null) ? eol : System.getProperty("line.separator"); namespaces = new HashMap(); } @@ -195,28 +121,17 @@ serialize(node, out, false); } - void serialize(Node node, final OutputStream out, + void serialize(final Node node, final OutputStream out, boolean convertToCdata) throws IOException { - while (node != null) - { - Node next = node.getNextSibling(); - doSerialize(node, out, convertToCdata); - node = next; - } - } - - private void doSerialize(final Node node, final OutputStream out, - boolean convertToCdata) - throws IOException - { if (out == null) { throw new NullPointerException("no output stream"); } String value, prefix; Node children; + Node next = node.getNextSibling(); String uri = node.getNamespaceURI(); boolean defined = false; short nt = node.getNodeType(); @@ -237,12 +152,7 @@ { break; } - String name = node.getLocalName(); - if (name == null) - { - name = node.getNodeName(); - } - define(nsuri, name); + define(nsuri, node.getLocalName()); } else if (uri != null && !isDefined(uri)) { @@ -260,8 +170,7 @@ out.write(encodeText(a_nodeName)); String a_nodeValue = node.getNodeValue(); if (mode == Stylesheet.OUTPUT_HTML && - a_nodeName.equals(a_nodeValue) && - isHTMLBoolean((Attr) node, a_nodeName)) + a_nodeName.equals(a_nodeValue)) { break; } @@ -418,18 +327,11 @@ for (Node ctx = html.getFirstChild(); ctx != null; ctx = ctx.getNextSibling()) { - if (ctx.getNodeType() == Node.ELEMENT_NODE) + if (ctx.getNodeType() == Node.ELEMENT_NODE && + "head".equalsIgnoreCase(ctx.getLocalName())) { - String name = ctx.getLocalName(); - if (name == null) - { - name = ctx.getNodeName(); - } - if ("head".equalsIgnoreCase(name)) - { - head = ctx; - break; - } + head = ctx; + break; } } if (head == null) @@ -459,39 +361,32 @@ for (Node ctx = head.getFirstChild(); ctx != null; ctx = ctx.getNextSibling()) { - if (ctx.getNodeType() == Node.ELEMENT_NODE) + if (ctx.getNodeType() == Node.ELEMENT_NODE && + "meta".equalsIgnoreCase(ctx.getLocalName())) { - String name = ctx.getLocalName(); - if (name == null) - { - name = ctx.getNodeName(); - } - if ("meta".equalsIgnoreCase(name)) + NamedNodeMap metaAttrs = ctx.getAttributes(); + int len = metaAttrs.getLength(); + String httpEquiv = null; + Node content = null; + for (int i = 0; i < len; i++) { - NamedNodeMap metaAttrs = ctx.getAttributes(); - int len = metaAttrs.getLength(); - String httpEquiv = null; - Node content = null; - for (int i = 0; i < len; i++) + Node attr = metaAttrs.item(i); + String attrName = attr.getNodeName(); + if ("http-equiv".equalsIgnoreCase(attrName)) { - Node attr = metaAttrs.item(i); - String attrName = attr.getNodeName(); - if ("http-equiv".equalsIgnoreCase(attrName)) - { - httpEquiv = attr.getNodeValue(); - } - else if ("content".equalsIgnoreCase(attrName)) - { - content = attr; - } + httpEquiv = attr.getNodeValue(); } - if ("Content-Type".equalsIgnoreCase(httpEquiv)) + else if ("content".equalsIgnoreCase(attrName)) { - meta = ctx; - metaContent = content; - break; + content = attr; } } + if ("Content-Type".equalsIgnoreCase(httpEquiv)) + { + meta = ctx; + metaContent = content; + break; + } } } if (meta == null) @@ -529,7 +424,6 @@ DocumentType doctype = (DocumentType) node; out.write(BRA); out.write(BANG); - out.write(encodeText("DOCTYPE ")); value = doctype.getNodeName(); out.write(encodeText(value)); String publicId = doctype.getPublicId(); @@ -574,6 +468,10 @@ { undefine(uri); } + if (next != null) + { + serialize(next, out, convertToCdata); + } } boolean isDefined(String uri) @@ -600,50 +498,37 @@ } final byte[] encodeText(String text) - throws IOException + throws UnsupportedEncodingException { - encoder.reset(); - if (!encoder.canEncode(text)) + if (compatibilityMode) { - // Check each character - StringBuffer buf = new StringBuffer(); int len = text.length(); + StringBuffer buf = null; for (int i = 0; i < len; i++) { char c = text.charAt(i); - if (encoder.canEncode(c)) + if (c >= 127) { - buf.append(c); + if (buf == null) + { + buf = new StringBuffer(text.substring(0, i)); + } + buf.append('&'); + buf.append('#'); + buf.append((int) c); + buf.append(';'); } - else + else if (buf != null) { - // Replace with character entity reference - String hex = Integer.toHexString((int) c); - buf.append("&#x"); - buf.append(hex); - buf.append(';'); + buf.append(c); } } - text = buf.toString(); + if (buf != null) + { + text = buf.toString(); + } } - ByteBuffer encoded = encoder.encode(CharBuffer.wrap(text)); - int len = encoded.limit() - encoded.position(); - if (encoded.hasArray()) - { - byte[] ret = encoded.array(); - if (ret.length > len) - { - // Why? - byte[] ret2 = new byte[len]; - System.arraycopy(ret, 0, ret2, 0, len); - ret = ret2; - } - return ret; - } - encoded.flip(); - byte[] ret = new byte[len]; - encoded.get(ret, 0, len); - return ret; + return text.getBytes(encoding); } String encode(String text, boolean encodeCtl, boolean inAttr) @@ -743,20 +628,5 @@ throw new RuntimeException(e.getMessage()); } } - - boolean isHTMLBoolean(Attr attr, String attrName) - { - attrName = attrName.toLowerCase(); - Node element = attr.getOwnerElement(); - String elementName = element.getLocalName(); - if (elementName == null) - { - elementName = element.getNodeName(); - } - elementName = elementName.toLowerCase(); - Collection attributes = - (Collection) HTML_BOOLEAN_ATTRIBUTES.get(elementName); - return (attributes != null && attributes.contains(attrName)); - } - + }